Devourer

Devourer is a generic REST API client for Python 2.7 and 3.4+.

You can also subclass it to wrap a set of system calls, FFI or a messy dependency.

It features an object-oriented, declarative approach to simplify the communication. It depends on the brilliant requests package as the gateway to API server.

Basic usage

from devourer import GenericAPI, APIMethod, APIError


class TestApi(GenericAPI):
    posts = APIMethod('get', 'posts/')
    comments = APIMethod('get', 'posts/{id}/comments')
    post = APIMethod('get', 'posts/{id}/')
    add_post = APIMethod('post', 'posts/')

    def __init__(self):
        super(TestApi, self).__init__('http://jsonplaceholder.typicode.com/',
                                      None,  # this can be ('user', 'password')
                                             # or requests auth object
                                      load_json=True,
                                      throw_on_error=True
                                     )

api = TestApi()
posts = api.posts()
post = api.post(id=posts[0]['id'])
comments = api.comments(id=post['id'])
new_post_id = api.add_post(data={'userId': 1,
                                 'title': 'Breaking news',
                                 'body': 'I just got devoured.'})
try:
    post = api.post(id=new_post_id)
except APIError:
    print('Oops, this API is not persistent!')

The init function gives details so you don't need to repeat them elsewhere, enables parsing json responses and raising exceptions on error. You can also obtain raw string with load_json=False and silence errors getting None instead when they happen with throw_on_error=False.

When calling methods:

data and payload kwargs will be passed to requests call as data and json parameters.
all keyword arguments matching schema will be used in schema.
all other kwargs will be passed to requests call as query string parameters.

Async usage

Devourer supports asynchronous calls using concurrent.futures's ThreadPoolExecutor. API subclasses are created and used as usual, they just need to inherit from AsyncAPI instead of GenericAPI.

Async mode will work with threads. They can be system threads or gevent threads, but that requires monkey patching them.

class AsyncTestApi(AsyncAPI):
    posts = APIMethod('get', 'posts/')

api = AsyncTestApi()
posts = api.posts()  # Send HTTP request, but don't block. Returns a `concurrent.futures.Future`.
result = posts_r.result()     # Retrieve result, blocking if the request hasn't finished yet.

Installation

You can just pip install devourer.

Documentation

Feel free to browse the code and especially the tests to see what's going on behind the scenes. The current version of docs is available on http://devourer.readthedocs.org/en/latest/.

There's also an article on the thought process behind devourer on http://bujniewi.cz/devouring-an-api/ - it's a five minute read, but it could answer a few questions you might have.

Questions and contact

If you have any questions, feedback, want to say hi or talk about Python, just hit me up on https://twitter.com/bujniewicz

Contributions

Please read CONTRIBUTORS file before submitting a pull request.

We use Travis CI. The targets are 10.00 for lint and 100% for coverage, as well as building sphinx docs.

You can also check the build manually, just make sure to pip install -r requirements.txt before:

pylint devourer --rcfile=.pylintrc
coverage run --source=devourer -m devourer.tests && coverage report -m
cd docs && make html

Additionally you can check cyclomatic complexity and maintenance index with radon:

radon cc devourer
radon mi devourer

The target is A for maintenance index, C for cyclomatic complexity - but don't worry if it isn't met, I can refactor it after merging.